Asymmetric Hearing Loss in Adult Cochlear Implant Recipients: Results and Recommendations From a Multisite Prospective Clinical Trial

Abstract

Objective: A multisite clinical trial was conducted to obtain cochlear implant (CI) efficacy data in adults with asymmetric hearing loss (AHL) and establish an evidence-based framework for clinical decision-making regarding CI candidacy, counseling, and assessment tools. Study hypotheses were threefold: (1) 6-month postimplant performance in the poor ear (PE) with a CI will be significantly better than preimplant performance with a hearing aid (HA), (2) 6-month postimplant performance with a CI and HA (bimodal) will be significantly better than preimplant performance with bilateral HAs (Bil HAs), and (3) 6-month postimplant bimodal performance will be significantly better than aided, better ear (BE) performance.

Design: Forty adults with AHL from four, metropolitan CI centers participated. Hearing criteria for the ear to be implanted included (1) pure-tone average (PTA, 0.5, 1, 2 kHz) of >70 dB HL, (2) aided, monosyllabic word score of ≤30%, (3) duration of severe-to-profound hearing loss of ≥6 months, and (4) onset of hearing loss ≥6 years of age. Hearing criteria for the BE included (1) PTA (0.5, 1, 2, 4 kHz) of 40 to 70 dB HL, (2) currently using a HA, (3) aided, word score of >40%, and (4) stable hearing for the previous 1-year period. Speech perception and localization measures, in quiet and in noise, were administered preimplant and at 3-, 6-, 9-, and 12-months postimplant. Preimplant testing was performed in three listening conditions, PE HA, BE HA, and Bil HAs. Postimplant testing was performed in three conditions, CI, BE HA, and bimodal. Outcome factors included age at implantation and length of deafness (LOD) in the PE.

Results: A hierarchical nonlinear analysis predicted significant improvement in the PE by 3 months postimplant versus preimplant for audibility and speech perception with a plateau in performance at approximately 6 months. The model predicted significant improvement in postimplant, bimodal outcomes versus preimplant outcomes (Bil HAs) for all speech perception measures by 3 months. Both age and LOD were predicted to moderate some CI and bimodal outcomes. In contrast with speech perception, localization in quiet and noise was not predicted to improve by 6 months when comparing Bil HAs (preimplant) to bimodal (postimplant) outcomes. However, when participants' preimplant everyday listening condition (BE HA or Bil HAs) was compared with bimodal performance, the model predicted significant improvement by 3 months for localization in quiet and noise. Lastly, BE HA results were stable over time; a generalized linear model analysis revealed bimodal performance was significantly better than performance with a BE HA at all postimplant intervals for most speech perception measures and localization.

Conclusions: Results revealed significant CI and bimodal benefit for AHL participants by 3-months postimplant, with a plateau in CI and bimodal performance at approximately 6-months postimplant. Results can be used to inform AHL CI candidates and to monitor postimplant performance. On the basis of this and other AHL research, clinicians should consider a CI for individuals with AHL if the PE has a PTA (0.5, 1, 2 kHz) >70 dB HL and a Consonant-Vowel Nucleus-Consonant word score ≤40%. LOD >10 years should not be a contraindication.

Figure 1.

Poor ear outcome curves show model-predicted performance as a function of time since the pre-implant test interval (circle). Median time points for the 3m, 6m and 12m intervals are represented by the triangle, square and diamond, respectively. Panels A, B, and C show outcomes for audibility, CNC60, and AzB60, respectively. Vertical lines indicate the first significant change in performance from pre-implant (solid line) and when performance begins to plateau (dashed line).

Figure 2.

Effect of age on model-predicted PE AzB60 performance. Dark and light gray shading represent 95% confidence bands for younger participants (Mean, −1 SD) and older participants (Mean, +1 SD), respectively.

Figure 3.

Effect of LOD on model-predicted PE audibility (Panel A), CNC60 (Panel B), and AzB60 (Panel C). Dark and light gray shading represent 95% confidence bands for participants with LOD < 10 years and those with LOD > 10 years, respectively.

Figure 4.

Bimodal outcome curves showing model-predicted performance as a function of time since the pre-implant test interval (circle). Median time points for the 3m, 6m, and 12m test intervals are represented by the triangle, square, and diamond, respectively. Vertical lines indicate the first significant change in performance from pre-implant (solid line) and when performance begins to plateau (dashed line).

Figure 5.

Effect of age on model-predicted bimodal outcomes. In Panels A-C, dark and light gray shading represents 95% confidence bands for younger (Mean, −1 SD) and older participants (Mean, +1 SD), respectively.

Figure 6.

Group mean scores for the generalized linear model analysis are shown for all test measures and intervals for the BE HA (light gray line) and Bil HA/bimodal (black line) listening conditions. The number (n) of participants for each measure is provided at the top of each panel. Asterisks indicate significance levels.